Standards for interoperability

From Stellar Deliverable 1.1

The integrated use of TEL systems knowledge and contexts is still a complex and rarely well implemented scenario which needs further research. Representing knowledge in an interoperable manner among various TEL systems is a key element. Current user centred standards for usability and accessibility have a strong orientation towards addressing the modelling of user interfaces and devices.

Interoperability for TEL has been mainly developed concerning instructional design and resources, encompassing tools and roles. Within the community several specifications/standards of content exchange are used that allow for exchange of learning content between different platforms. SCORM (Sharable Content Object Reference Model) was one of the first standards to be used for TEL systems interoperability. Its first version was mainly focused on content aggregation, the last one on activity sequencing on content objects (2004). LOM (Learning Object Metatada) was been set up in 2002 to describe and share learning objects within a LMS. Whereas LOM represented a first approach, no interoperable representation of domain elements was provided as the classification category left open the issue of an interoperable classification system. A semantic Web approach provides an interoperable language (OWL) with a well-founded semantics that could be used to provide ontologies for describing content element in educational systems. There is a suggestion that we can usually find what we want (on the Internet). We have good search engines, so why is there a concern about interoperability? The argument is that search engine technologies are based on natural language, and while there is a recognition that they are usually good, they fail in some respects, for example if you search for a vehicle with two wheels in a natural language search engine, it is unlikely that (in the present state of the art) ‘bicycle’ will be returned. On the other hand, the semantic web is based on metadata which is concerned with providing computable semantics to data. A search using semantic web technologies would return ‘bicycle’ in the example above.

Representing competences is a way to solve the issue of annotating resources with a related domain content. More recently this has been tackled by a variety of projects, but some essential problems such as understanding the ways in which people work with competencies on a large scale and how to generate metadata easily still remain. For example the competencies used by the PÏSA studies (OECD) regarding mathematics, science and reading are very generic ones and are directly instantiated into questions and not into precise competencies. IMS Reusable Definition of Competency or Educational Objective Specification is only a first step in competence interoperability, providing only textual descriptions, no computable semantics and no way of relating competencies to each other.

TEL systems and resources are now integrated into larger environments (contexts) and used outside the classroom. As such contexts play an important role in supporting the learning process, and interoperable representation of context becomes essential.

As recent projects in the e-content plus program applied in several application domains, key issues currently being researched include:Federation of distributed and fragmented content resources; Federation of existing content repositories via for example LOM application profiles, and harvesting and publishing protocols as OAI-PMH; Mapping of varying metadata formats and interpretations as also the development of a shared understanding and usage of different types of metadata, so as competence metadata (IMS-RDCEO); Integrated use of different classification and descriptions formats on competences, domains, usage metadata, and context metadata; Enrichment of federated repositories in active education usage as also the integration of metadata usage in instructional designs using it for "finding content"; Access to, and findability of, content, based on user-driven needs and intuitive visualisations; Sensemaking and usage of standards in PLE and web 2.0 driven learning environments as also mash ups.

There is also a need for standardisation in sensor networks. These should be interoperable but there are no standards and this means that different sensor networks work in different ways and hence cannot work together to realise the benefits of all networks. How could these be standardised and what might the implications of standardisation be for education?

Research questions include:

• For education settings, what do we think would be useful if interoperability were improved?

• Delivery in real time is a challenge (e.g. yahoo pipes). What is required in order to be able to achieve this? How would this enhance learning?

• Consider the idea of sharing resources and the idea of shifting context. Learners with mobile devices can move between contexts. What are the different aspects of context (e.g.location)?

• Why is it important to make different contexts interoperable?